论文部分内容阅读
本文提出了一种改进型的经验模态分解算法用于心音图(PCG)信号去噪,结合PCG的规则平均Shannon能量包络算法,可有效提取PCG中的S1/S2成分。首先,通过小波变换和经验模态分解结合算法对PCG信号进行滤波预处理;然后,提取预处理后PCG信号的固有模函数(IMF)时域、频域特性及能量包络;最后,结合信号的Shannon能量包络和IMF相关特性准确定位出S1和S2。运用该方法对30例PCG信号进行测试,得到S1/S2成分的综合识别率达99.75%。实验结果表明,本文算法运用于S1/S2成分提取具有较好的效果,为进一步研究心音身份识别奠定基础。
In this paper, an improved empirical mode decomposition algorithm is proposed for PCG signal denoising. Combined with PCG regular Shannon energy envelope algorithm, S1 / S2 components in PCG can be effectively extracted. Firstly, the PCG signal is filtered and preprocessed by combining the wavelet transform and the empirical mode decomposition. Then, the intrinsic mode functions (IMFs) of the preprocessed PCG signal are extracted in the time domain, the frequency domain and the energy envelope. Finally, the combined signal Shannon energy envelope and IMF-related features accurately locate S1 and S2. Using this method to test 30 PCG signals, the comprehensive recognition rate of S1 / S2 components was 99.75%. The experimental results show that the algorithm proposed in this paper has good effect on S1 / S2 component extraction and lays the foundation for further research on vocal identification.